SPEAKER: In this episodeof "BigQuery Spotlight,"we're going to dive deepinto how to ask a questionor run a query in BigQuery.Running queries is one ofthe most fundamental partsof discovering insightsfrom your data.Today's question--what is the best jerseynumber you should choosein order to improveyour basketball game?Stay tuned and find out.[MUSIC PLAYING]It's no surprisethat big data hasmade its way into the worldof professional sports.Teams at everylevel are startingto gather, process,and analyze dataso that they can get themost out of their playersand find a competitiveedge in the game.To see how this wouldwork, let's check outthe NCAA basketball publicdata set available in BigQuery.The data set containsplay-by-play datafor several years'worth of games.You can get information likefouls, free throws, scores,player numbers, timeouts,basically as much dataas the pros have.So let's see if we can usethis data for today's questionand determine whichjersey number is the bestthree-point shooter.If you want tofollow along, clickon the link in thedescription belowto set up your ownBigQuery sandboxwhere you can analyze datawithout needing a credit card.Once you're in theBigQuery sandbox,head over to the GoogleCloud Marketplaceby clicking on Add Dataand then Explore Data Sets.Search for the NCAAbasketball and clickto launch the data setin the BigQuery UI.The NCAA basketball data set hasa number of different tables.Clicking on theplay-by-play table,you can first viewthe table details,which indicates thatthere are over fourmillion rows of events.That's a lot of basketball data.Check the tableschema, and you'llsee each column or field ofdata that's available for eachof those events.But first, let's look atwhat those events are.You can start toexplore the databy clicking on thePreview tab to viewa few rows of sample data.But to get a more comprehensivelist of all the eventtypes, the best way tofind out is to run a query.Click Query Tableand a blank querywill appear in the editor witha table reference pre-populated.Before we write our query,it's important to knowthat the language used tocommunicate with BigQueryis SQL, or StructuredQuery Language.SQL is also thestandard languagefor communicating withrelational databases.BigQuery is ANSI SQL compliant,so experienced practitionerswill hit the ground running.Don't worry ifyou're not SQL savvy.There are plenty of resourcesfor learning or brushingup on your SQL skills justa quick web search away.So now, let's copyin a SQL queryto explore the event types.The key command in SQLfor retrieving datais SELECT, whichwe'll use to pullin a list of all of thedifferent event typesand number of occurrencesfrom the play-by-play table.Notice that the table referenceis in the format of the GCPproject name, thendata set name,then table name,separated by dotsand enclosed with backticks, not quotes.We'll select all of the eventtypes and group the resultsso that we can get thecount of each event.Once the SQL query is written,click on the green check markon the right-handside of the windowto open the query validator.A green check markmeans the query is validand will show theestimated amount of datathat the query willprocess when you run it.If the query is invalid, ared exclamation point iconis displayed.You can click on itand get some guidanceon how to correct the problem.Click Run andBigQuery gets to work.After the queryexecution is complete,the query servicedisplays the resultsas a table in the web UI.You can see that theEvent Type column hasboth three point made andthree point miss events.Now we can writea new query whichselects the jersey numberfor each play-by-play eventand then uses theEvent Types columnto determine thethree point field goalaccuracy for each jersey number.Click Run and then in less thana second we have the answer.Looks like number 45 sinksthe highest percentageof their three-pointer attempts.Of course, buying anew jersey doesn'tmake you a betterbasketball player, right?An important note-- we cansee that the table we queriedhas three gigabytes of data.But BigQuery only read thecolumns needed and thereforejust processed 78megabytes of data.Well, basketballenthusiasts out there,try out the NCAAdata set for yourselfand ask questions todiscover what you mightbe able to do to up your game.If you want to learnmore about BigQuery,check out the restof our playlist.And don't forget tosubscribe to our channelon YouTube forother great content.Look out for the next episodeof "BigQuery Spotlight."And remember, stay curious.[MUSIC PLAYING]
